search

siat-nlp / TransDG

Source code for the paper "Improving Knowledge-aware Dialogue Generation via Knowledge Base Question Answering".
MIT License
64 stars 15 forks source link

No such file or directory: './data/Reddit/train_v3.txt #4

Open hackbuteer001 opened 3 years ago

hackbuteer001 commented 3 years ago

Hello, Thanks for the code. when i run "sh gen_linkings.sh" , No such file or directory: './data/Reddit/train_v3.txt

I got "Reddit" from "http://coai.cs.tsinghua.edu.cn/file/commonsense_conversation_dataset.tar.gz" and "train_v3.txt" was not been found. Is there anything wrong with the download link?

iwangjian commented 3 years ago

Hi, sorry for the delayed feedback. The train_v3.txt should be named train.txt instead, which is formed after retrieving top-k similar responses. I have fixed the bug and updated the README. Thanks for your comment.

CYAugust commented 3 years ago

Hello,according to the original author's link, I couldn't find the Reddit data set. I got "commonsense_conversation_dataset. tar.gz" from your link but couldn't find the "csk_entity.txt" file. Is this file in ”data.tar.gz“ of ”commonsense_conversation_dataset. tar.gz“? I don't know why I can't unzip this file in any way.

iwangjian commented 3 years ago

Sorry for the ambiguous instructions in README, the "csk_entity.txt" is the content of the key field "csk_entities" in the "resource.txt" extracted from the tar.gz file, where each line is an entity. Similarly, the "csk_triples.txt" is the content of the key field "csk_triples" in the "resource.txt".